You are at The Krib ->Plants ->CO2 [E-mail]

DIY CO2 Injection: The Yeast Method

By Thomas Narten


This article gives instructions for a cheap Do-It-Yourself CO2 injection system. The CO2 is produced by a mixture of sugar, yeast and water, and the setup is constructed entirely from cheap and readily available materials. Thanks go to Brian R Silverstein ( for posting a description of his setup, which initially used sourdough starter. His message got me started and much of the information below comes from his postings. Not surprisingly, Brian thanks Ulli Kaeufl ( & George Booth (booth-at-hplvec.LVLD.HP.COM). Finally, a big thanks go to the many participants of the Usenet *.aquaria newsgroups, where this topic has been discussed many times, and numerous persons report positive results using this system.

Warning: This setup may not be suitable for everyone. It requires regular fiddling and the uncertain CO2 injection rate may lead to pH fluctuations that can kill fish (especially if you have soft water). My water is fairly hard (GH 140-180 ppm) and the pH varies from 7.8 with no CO2 to 7.0 with CO2. Read the FAQ on plants and CO2 injection before attempting this. CO2 is an important contributor to robust plant growth, but it is advisable to understand the big picture before going this route. It is particularly important to understand the relationship between pH, CO2 concentration and carbonate hardness. Note also that CO2 injection is a complete waste of time if one has inadequate light, as is typically the case with single-tube strip lights sold at most aquarium stores. One bulb usually produces an insufficient quantity of light, and typical bulbs do not produce the "full spectrum" of light plants require.

On the other hand, the described set up is an inexpensive way to observe first hand the benefits of CO2 injection on plant growth. I personally had never seen what CO2 could do for a tank before trying it myself, and would probably never have purchased a conventional CO2 injection system because of the high cost. In fact, I still haven't purchased one. I find the yeast method adequate for my current needs. I have been using it now for 8 months and have managed to acquire and successfully grow over 25 species of plants with only a handful of failures. Here are the steps.


The first step is to create a CO2 generator. I use a 2-liter plastic soft-drink container. The "hard" part is attaching a hose to the container so that the CO2 doesn't leak out on its way to your aquarium. I use a number 3 rubber stopper with a 1/4 inch hole in the middle. A short segment of 1/4 inch copper tubing goes through the hole and connects to regular air hose tubing. I bought a stopper and copper tube from a campus chemical supply store for $ .45.

Note: Other folks have had success attaching the tube directly to the resealable cap. Simply poke or drill a hole through the cap, stick the tubing through it and seal it with the aquarium silicone sealer. The variations are endless.


The second step is to create a CO2 reactor for trapping and dissolving the CO2 bubbles before they escape to the surface of your aquarium. Note: For the first few weeks, I simply placed an airstone at the bottom of the tank and let the CO2 bubble to the surface. While inefficient, enough CO2 dissolved to make my plants grow noticably faster. Use of an airstone is suggested because the increased surface area of many small bubbles helps dissolve the CO2 faster.

More recently, I constructed a "sophisticated" reactor. The idea is to trap the rising CO2 bubbles before they reach the surface and escape. Doing so provides two benefits. First, less escaping CO2 means more CO2 in your tank. Second, the trapped CO2 bubbles act as a (very!) crude gauge of how much CO2 you have in your tank. If a lot of CO2 is trapped, the amount of dissolved CO2 in your tank is likely to be high. If little CO2 is trapped, your rig may not be producing CO2 at a high enough rate to replace that used by plants or escaping from your tank.

My reactor consists of a small clear plastic bowl that flower pots are placed on ($0.40 at any gardening store). I cut it in half, glued a clear piece of plastic to it and wedged it between the filter tube and aquarium glass just below one of the powerheads. The cutaway portion faces down creating a trap for rising CO2 bubbles. It is best to use clear plastic so that you can see how much CO2 is accumulating in the reactor.

Other folks have had success with a more sophisticated reactor that actually circulates water in the reactor in order to dissolve the CO2 faster. A nice reactor can be made from the wide end of a gravel cleaner. The 1/2 inch tubing attached to such cleaners connects directly to the exit of many powerheads. The reactor catches the CO2 bubbles, while the water from the powerhead churns and dissolves the CO2 and then flows into the tank.

Tetra also markets a "CO2 Bell", a simple reactor with suction cups that attach to the side of the tank. Since it is ready to go, it is the easiest to use, but finding them at local stores may be problematical, since few stores seem to market CO2 injection related products.

If you use a cannister filter, another possibility is to let the CO2 bubble into the filter intake. The CO2 will completely dissolve by the time it exits the filter. The "gravel cleaner" reactor described above can be used to catch the CO2 with the hose connecting it to the filter intake.


The third step is to create CO2. I mix 3/4 cup sugar, 6 cups water and a pinch of yeast (roughly 1 teaspoon does the trick). The exact ratio of ingredients is not critical. For example, I use enough water to fill the bottle to about 3 inches from the top. You don't want to get the mixture too full, as the mixture may foam a bit, and you don't want the foam flowing into your tank.

Using the above mixture, I regularly get 10-14 days of strong CO2 production, at a rate of one bubble every 3-7 seconds. When the system stops bubbling (the drop in gas production is often precipitous), dump out everything and mix up a new concoction.


1) I doesn't hurt to shake the mixture every 2-3 days. Early on it seemed to me that a mixture produced CO2 longer if I shook it every couple of days or so, but I am no longer sure that this is actually the case. (Note: it will be difficult to shake the mixture if it is too full!).

2) Hot water contracts when cooled. Actually, the hot air trapped in your rig contracts when cooled. If your mixture is warmer than room temperature, it may suck water out of your tank before it starts producing enough CO2 to reverse the flow. If your CO2 rig sits below your aquarium, bad things could happen.

One way to prevent problems is to insert a standard air valve between the mixture and the airstone. Right after mixing up a new batch, close the valve. Once the rig is producing CO2 again, open the valve.

Rather than bother with a valve, I simply keep a spare 2-liter bottle on hand. When the current batch looks like it's starting to peter out (or before, if you don't want to have your tank's pH go up as CO2 production wanes), mix up a new batch, but don't connect it right away. Let it sit in a corner for 12-24 hours and then hook up the new mixture when you know it's going strong. That way there is no interruption in CO2 flow through your tank.

Another possibility is to connect two separate mixtures together using gang valves (e.g., use a T-connector to attach both mixtures to the same airstone). Then, mix up a new batch in one rig while the other one is still going. The idea is to always have one of the mixtures going at full steam so that when one mixture quits, the other produces enough CO2 to keep your plants happy.

3) Hot water kills yeast. If you dissolve yeast in hot water (like what comes out of my gas water heater), you probably won't get any CO2 from your setup. If you've mixed things up properly, you should get CO2 production within a few hours, possibly more quickly; I've sometimes had things going 30 minutes after mixing a new batch. Here's what I do:

  1. Fill a 2 cup measuring cup with hot water, and dissolve the sugar in it. If you use a cup with a spout, you can pour the mixture directly into the soda bottle without a funnel (one less gadget to clean!)
  2. Add three more cups of cold water to the mixture and shake it up.
  3. Dissolve the yeast in a cup of cold water and add to mixture. I've only used bakers yeast, but others should work too.
  4. Add more cold water to mixture to fill to desired level. Shake the mixture well.
  5. Let sit at room temperature for a few hours until the mixture reaches room temperature. That's often all the time that is needed to get initial CO2 production going. I feel safer connecting my setup to the tank *after* I know that it is producing gas and won't suck water out of the tank.
4) Some batches of yeast take longer get going than others. For example, yeast stored in a refrigerator takes longer to become active than yeast stored at room temperature. Letting the yeast sit out at room temperature overnight before adding it to your mixture reduces the startup delay. Another suggestion (from homebrewers) is to hydrate the yeast in a cup of warm (80 degree) water for 20 minutes before adding it to the sugar mixture.

5) Nicholas Plummer reports that most people who advocate using a cannister filter as a reactor seem to have Eheims. When he tried bubbling gas into the intake of a fluval 203, the gas made the impeller rattle loudly when a bubble formed at the top of the cannister. He suspects that the slightly poorer design of the Fluval is to blame.

Even without funneling CO2 into the filter, he gets the same effect (but to a lesser degree) late in the day. The water becomes so saturated with oxygen from photosynthesis that it starts to bubble out in the cannister. It reenters solution during the night and is always completely gone by the next morning. Perhaps the slight agitation of the water entering the cannister through an eheim surface extractor causes oxygen to precipitate out of solution.

7) Some folks appear tempted to turn off the CO2 production at night and back on again in the morning. After all, plants consume oxygen at night and produce more CO2. This is a mistake. It is better to keep the tank's CO2 concentration relatively constant because a drop in the dissolved CO2 concentration leads to an increase in pH. Neither plants nor fish seem to appreciate rapid pH swings.

8) Several persons report the appearance of an apparently harmless white film on their CO2 airstones. I have not witnessed this in my tank and am not sure of its cause.

9) I am still experimenting with exact ratios of sugar, yeast and water. In my system, the rate of CO2 production is relatively steady for a number of days, but then drops precipitously (from several bubbles a minute to no production in 24-48 hours). I suspect that the yeast exhausts some trace element or other critical ingredient. Yeast needs a number of elements, including nitrogen (e.g., ammonia but not nitrate), magnesium, phosphorous, potassium, calcium, zinc, iron and copper. My initial hypothesis was that a buildup of alcohol limited yeast production. However, a homebrewing friend reports that it would take roughly 3.75 cups of sugar in a 2-liter bottle to produce enough alcohol to reach a concentration of 15%, the point at which no strain of yeast can survive. As another data point, I have been running a mixture now for over six weeks that is still producing an adequate amount of CO2. I have no explanation for the longevity of the current mixture. I used the same general recipe that previously worked for only two weeks.

10) It should be possible to culture a new mixture using yeast from a previous batch. Simply save the sludge that collects at the bottom of a mixture. The top layer of sludge, which should be a bit lighter in color, consists of live but dormant yeast cells. The darker layer underneath is dead cells. I have not tried this myself. For more information, look towards the homebrewing field.


After some 15 years of mixed success at growing plants, I now have a jungle. I have to trim the growth in my 55g tank every 1-2 weeks. If I don't, plants completely cover the surface of the tank preventing light from reaching the lower levels. I have so many plants I'm throwing them away! My setup consists of: Without the CO2 rig, the measured CO2 concentration in my tank is about 1-2 ppm (according to the Lamotte CO2 test kit). With the CO2 rig, measured concentration ranges from 5-20 ppm, depending on what my bubble rate is. With CO2 injection, my pH drops from 7.8 to about 7.0 (at a CO2 concentration of 20 ppm). When the CO2 runs out, pH rises again. Warning: The pH change can happen in less than 24 hours, and some fish may suffer. To avoid this problem, hook up a new mixture before the old one has completely run out.

I recently counted over 25 species of plants in my 55g tank. About 15 species I've had for eight months, the rest for about four. Most of the plants are doing extremely well. My major successes include hornwort, vallisneria, pennywort, elodea, foxtail, wisteria, ludwigia, java fern, rotala rotundifolia and both hygro. polysperma and carymbosa, which I throw out regularly.

Other successes include willow moss, several swords, various cryptocorynes, alternanthera lilacinia, rotala macrandra, sagittaria, plus a few things I haven't been able to identify. My anubia nana doubled its size in four months, more growth than the entire previous year. It's growing a new leave every week. Last month, it bloomed for the first time.

Some of the plants are not doing much, but they haven't died and rotted either. Part of the problem is that I don't have enough light reaching the bottom. The faster growing plants outcompete everything else. I have also witnessed a number of plants near the bottom struggling weeks to grow a few inches (even when under direct light), but after reaching a height of four or five inches, growth takes off. I'm convinced now that more light would help.

My only true failure is cabomba. I had it growing very well for 2-3 months (I was throwing it away I had so much), but then growth just stopped. The plants got stringy looking and the leaves fell off. Since then I have learned that I probably don't have enough light for cabomba in my tank and my water is warmer (78 degrees) than ideal.

It is now apparently also the case that my tank doesn't contain enough nitrogen to keep all plants growing at full steam. In the three months since purchasing a sensitive Lamotte nitrate test kit, nitrate levels have always registered zero. It is also apparent that some plants are growing at the expense of others. For example, elodea and foxtail growth was prolific at first, but now grows only slowly.


In summary, the described sugar, yeast and water CO2 injection system provides an inexpensive way to experiment with CO2 in a plant tank. It is particular attractive to those interested in trying CO2, but skeptical that benefits justify the cost of conventional systems. Although a bit more hassle to maintain than a proper tank-driven injection system, I do not find the burden to be excessive. Although I could afford to purchase the more expensive conventional system, I have little inclination to do so at this point. I spend much more time feeding my fish and trimming plants than I do fiddling with my rig.
Thomas Narten
16 Leto Rd.
Albany, NY 12203

Hypertext translation by E. Olson from Revision 3, 4/22/94

Note: This document can be freely redistributed to other mailing lists, reprinted in local fish club bulletins, etc. as long as it is reprinted in its entirety and proper credit is given. If you want to reprint it, please contact me first in order to get the latest version. Further comments and suggestions are always appreciated.

Revision 3 of this article appears in the May-June 1994 issue of "The Aquatic Gardener" (TAG). TAG is a publication produced by the Aquatic Gardening Association, which is dedicated to the dissemination of information concerning the growing of aquatic plants. Membership includes the bi-monthly TAG journal. For a one-year membership, send $15 (US dollars) to:

Dorothy Reimer
83 Cathcart St.
London, Ontario
Canada N6C 3L9
I strongly encourage plant enthusiasts to join this organization. I have found a lot of practical information on fertilizers, CO2 injection, etc in TAG. The best thing about TAG is that its contributers have lots of experience and success growing plants. Back issues are $3 each, and I think few people will disappointed at getting the last 1-2 years worth of back issues. Back issues are $3 each and can be ordered from

Diana Walstad
2303 Mt. Sinai Rd.
Chapel Hill NC 27514
Up to CO2 <- Plants <- The Krib This page was last updated 29 October 1998